My 2 cents - releasing the connections seems to be related the to wait_timeout setting in MySQL, which is by default 28800 seconds (8 hours) [1]. By repeatedly running the integration tests the max_connections are being exhausted quickly and the issue may not have been caught before when running them over a longer period of time. It may be useful to try setting that to a lower value.

Source: [1] https://dev.mysql.com/doc/refman/5.0/en/server-system-variables.html#sysvar_wait_timeout

-Vaikunth

On Tue, May 5, 2015 at 1:35 AM, Jacek Becla <[log in to unmask]> wrote:

Guy, we have a pretty major mysql connection leak, see

https://jira.lsstcorp.org/browse/DM-2698

before you complain about any strange behavior
or random crash of Qserv.

I will try to get to the bottom of it, if anyone has
any quick suggestions where the problem might be, you
are welcome to jump in or take over.

Jacek

########################################################################
Use REPLY-ALL to reply to list

To unsubscribe from the QSERV-L list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=QSERV-L&A=1